CN104041147A - Power management in a cellular system - Google Patents

Abstract

A method of controlling the respective transmit powers allocated by a base station of a cellular communications network to each of a plurality of sub-bands is provided. Information from at least one other base station of the network is received, the information comprising information about a sensitivity of a utility function in a cell served by the other base station to changes in powers allocated to respective sub-bands by the base station. A sub-band is identified in which it would be relatively advantageous to increase a transmit power. It is determined whether a factor relating to a happiness of users in the cell exceeds a threshold value. The transmit power in the identified sub-band is increased only if the happiness factor is less than the threshold value.

Description

Power management in cellular system

The present invention relates to cellular phone network, and relate more specifically to the method for the power of the signal that transmits the base station for controlling in such network, to reduce the disturbing effect of such transmission, safeguard the performance of desired network simultaneously.The invention still further relates to the base station in such network.

Compare with third generation system, such as Long Term Evolution (LTE) the 4th generation (4G) cellular system be currently developed to improve systematic function and user data rate.System is although it is so designed to improve systematic function and user data rate, but Special attention will be given to is the systematic function of booster injection to the user at cell edge.Realizing so improved one of the most effective mode is by power and interference management.

Although power and interference management are designed to come raising system and user performance by reducing unnecessary interference originally, importantly recognize that this can realize by reduce as much as possible delivering power when still meeting certain satisfaction target.By removing unnecessary delivering power, improving to a large extent energy efficiency, be possible.Although for single base station, energy efficiency may not be a serious problem, is height correlation in the situation that large-scale base station network is deployed this problem.

In typical deployment scenario, community is not individualism, this means community institute that each community may be close to around.Thereby, when mobile subscriber away from serving BS when move contiguous community, call quality reduces, this is not only because the weakening of serving BS signal, and because from the increase of the interference of the community of one or more main vicinities.Such interference is commonly called presence of intercell interference, and alleviating of such interference is considered, and to promote the user of Cell Edge User, experiences.Interference management for LTE is more complicated than the interference management in the old-fashioned 3G system such as wideband code division multichannel access (WCDMA) system, because LTE system relates to the power division in time and frequency domain, and WCDMA system only relates to the distribution of time-domain.

A kind of known method that alleviates presence of intercell interference is by use, to be called as the method for partial frequency multiplexing (FFR), wherein the mobile subscriber at each center of housing estate place is assigned with same frequency, and the user of cell edge is assigned with the frequency subsets different from those users of edge of community next-door neighbour.Therefore, presence of intercell interference at cell edge place can significantly be reduced (R.Kwan, C.Leung, " A Survey of Scheduling and Interference Mitigation in LTE (investigation that the scheduling in LTE and interference suppress) ", volume 2010, article ID273486).

Although FFR and variant thereof are that they are tired out by following shortcoming for disturbing the known technology of inhibition: the frequency subsets that is used to cell edge mobile subscriber need to be by careful planning, and this planning is normally carried out statically in the network planning stage.Thereby such method is also not suitable for the Femto cell (femtocell) that wherein base station is disposed by specific (ad hoc) mode.In addition, these methods are not considered dynamic subscriber's flow distribution, and thereby have reduced the availability of frequency spectrum.

On the other hand, can be by making the monthly dynamics of power and frequency resource change (D.L ó pez-P é rez by the mode of centralization for each user's co-allocation frequency, power, modulation and encoding scheme (MCS) in community, G.de la Roche, A.Valcarce, A.J ü ttner, J.Zhang, " Interference Avoidance and Dynamic Frequency Planning for WiMAX Femtocells Networks (for the interference of femto network, avoiding and dynamic frequency planning) ", Proc.of ICCS, 2008).But such method needs centralized entity, and computational complexity is unrealisticly high.

At A.L.Stolyar, H.Viswanathan, " Self-organizing Dynamic Fractional Frequency Reuse for Best-Effort Traffic Through Distributed Inter-cell Coordination (by coordinating between distributing cell that optimal flux is carried out to self-organizing dynamic part channeling) ", proc.of IEEE Infocomm, in 2009 4 months, a kind of algorithm based on gradient is suggested, and wherein frequency reuse mode is dynamically adapted to flow distribution.Because this method is to carry out self between community with distributed way, so do not need frequency planning.In addition, it is a kind of by the mode of distributed way dividing frequency that this method not only provides, and the method also allows power dynamically to be regulated by frequency, thereby the extra degree of freedom is provided.Although this method is useful, the document does not provide about service quality (QoS) how to be taken into account in the details in formulism.Therefore, power division may be not necessarily suited for serving institute's actual needs, thereby has reduced power efficiency.In addition, although the document provides the useful framework of dynamic disturbance inhibition aspect, about the problem of implementation aspect, still exist.For example, the analysis relation that this formula hypothesis knows for sure between spectrum efficiency and signal and interference and noise ratio (SINR).In fact, because different suppliers may have their receiver implementation and therefore have this fact of different performances, there is not such fixed relationship.

According to an aspect of the present invention, provide a kind of control by the base station assigns of cellular communications networks the method to each delivering power of each subband in a plurality of subbands, the method comprises:

Reception is from the information of at least one other base station of this network, and this information comprises that utilance function in the community about being served by this other base station is to the information to the susceptibility of the variation of the power of each subband by this base station assigns,

Identification improves the relatively useful subband of delivering power;

Whether the relevant factor of happiness (happiness) that judges the user in Yu Gai community surpasses threshold value; And

Only, in the situation that this happiness factor is less than this threshold value, just improve the delivering power in the subband identifying.

According to an aspect of the present invention, provide a kind of control by the base station assigns of cellular communications networks the method to each delivering power of each subband in a plurality of subbands, the method comprises:

Information from the first base station is sent to at least one other base station of this network, this information comprises that utilance function in the community about being served by this first base station is to the information to the susceptibility of the variation of the power of each subband by this other base station assigns,

Wherein the step of this transmission information comprises: by X2 interface, send information to this at least one other base station.

The method that control is given each delivering power of each subband in a plurality of subbands by the base station assigns of cellular communications networks, the method comprises:

Information from the first base station is sent to at least one other base station of this network, this information comprises that utilance function in the community about being served by this first base station is to the information to the susceptibility of the variation of the power of each subband by this other base station assigns,

Wherein the step of this transmission information comprises: transmit the information relevant to the relative arrowband delivering power of each subband Zhong Gai community in this subband.

According to an aspect of the present invention, provide a kind of method of determining the disturbing effect in the community served by the base station of cellular communications networks, the transmission of the base station of this interferences at least one community being close to from this network causes, the method comprises:

From being connected to the mobile device of this base station, obtain measurement result; And

Utilize this measurement result obtain about by this base station, served utilance function in Gai community to this other base station assigns measuring to the susceptibility of the variation of the power of each subband.

According to an aspect of the present invention, provide a kind of method of estimating the spectrum efficiency of the subband in the base station on cellular communications networks, the method comprises:

Power function by the CQI reported by the mobile device of measuring on this subband is similar to this spectrum efficiency;

By the linear function of the ratio of interference and noise being similar to this CQI by the measured signal of this mobile device, wherein this signal is measured by decibel to the ratio of interference and noise.

According to an aspect of the present invention, provide a kind of method of controlling the base station in cellular communications networks, the method comprises:

For each user in a plurality of users, receive the value of the initial bit rate requirement that represents this user;

Determine and be required that the corresponding downlink power of distributing to this user is to reach corresponding bit rate requirement;

Determine that total downlink power requires the summation into desired each downlink power; And

When total downlink power of base station surpasses threshold value, the bit rate of at least one user in these users is required to be reduced to the value requiring lower than corresponding initial bit rate.

According to an aspect of the present invention, provide the method for the value of the load on a kind of base station of calculating cellular communications networks, wherein this base station can utilize a plurality of subbands and can utilize frequency selectivity power to control, and the method comprises:

The value of the average power based on for each user and mean bit rate computational load.

According to an aspect of the present invention, provide a kind of control by the base station assigns of cellular communications networks the method to each delivering power of each subband in a plurality of subbands, the method comprises:

In this base station, from being connected to the mobile device of this base station, obtain channel quality information;

For each subband, utilize from the channel quality information formation average channel quality of this mobile device and measure; And

According to this average channel quality, measure, estimate about the utilance function in the community of being served by this base station the information to the susceptibility of the variation of the power of each subband by other base station assigns.

According to an aspect of the present invention, provide a kind of base station that is adapted to be the method for carrying out any other side.

For understanding better the present invention, and demonstration the present invention can how to be implemented, will to accompanying drawing, carry out reference by way of example now, in the accompanying drawings:

Fig. 1 show according to such as Long Term Evolution (LTE) the 4th generation the cellular communications networks that operates of (4G) cellular standards a part.

Fig. 2 shows the base station in the network of Fig. 1.

Fig. 3 is the figure that illustrates the effect of considering the happiness factor.

Fig. 4 is the figure that illustrates the effect of considering the modified happiness factor.

Fig. 5 is the flow chart illustrating according to the first method of the present invention.

Fig. 6 illustrates for the utilance of the various values of scale factor over time.

Fig. 7 illustrates for the power of the various values of scale factor over time.

Fig. 8 is the flow chart illustrating according to the second method of the present invention.

Fig. 9 illustrates for the utilance of the various values of scale factor over time.

Figure 10 illustrates for the power of the various values of scale factor over time.

Figure 11 illustrates the relation between ratio, CQI and the spectrum efficiency of signal and interference and noise.

Figure 12 further illustrates the relation between ratio, CQI and the spectrum efficiency of signal and interference and noise.

Figure 13 illustrates the available connection between the base station in may the disposing of Femto cell and macro cell base station.

The frequency that Figure 14 illustrates in may the disposing of Femto cell and macro cell base station is distributed.

Figure 15 illustrates the relation between utilance, power and desired bit rate.

Figure 16 is the second diagram of the relation between utilance, power and desired bit rate.

Fig. 1 show according to such as Long Term Evolution (LTE) the 4th generation the cellular communications networks 10 that operates of (4G) cellular standards a part.Network 10 is included as Macro base station or the enhancement mode Node B (eNB) 12,14 of community 16,18 services separately, be appreciated that between Zhe Liangge community 16,18 and have overlapping region, wherein subscriber equipment can with base station 12,14 in any connect.

16,18 a plurality of femto cell base stations or the family's enhancement mode Node B (HeNB) 20,22,24,26,28,30,32,34 that are built-in with the respective cell service of respectively doing for oneself in its vicinity in community.As everyone knows, in a macrocell, can there is tens of, hundreds of or or even thousands of Femto cell.Fig. 1 for the sake of clarity only shows a small amount of such Femto cell.For example, these Femto cells can be had by the user of cellular network respectively, or they can they be placed in place wherein (such as, market, campus, office garden or large-scale office building) common management under.

Fig. 2 illustrates in greater detail the formation of a base station in the base station in network.Base station 40 shown in Fig. 1 can be Macro base station or femto cell base station.

Base station 40 has transceiver circuit 42, for signal being converted to and being changed from passing through the needed form of air interface transmission.As the above mentioned, in this illustrated examples, base station is intended to form a part for LTE network, and therefore transceiver circuit is converted to signal and change from its needed form.Antenna 44 is connected to transceiver circuit 42.

Base station also has for being connected to the interface circuit 46 of the remainder of network.In the situation that base station 40 is femto cell base station, interface circuit 46 can for example be suitable for signal to be converted to and to change from be connected the needed form of transmission by broadband internet.In the situation that base station 40 is Macro base station, the dedicated link that interface circuit 46 can for example be suitable for signal being converted to and being changed Zi the core network by cellular communications networks transmits needed form.

Modulator-demodulator 48 is connected between transceiver circuit 42 and interface circuit 46, for the treatment of signal and therefrom extract related data.Modulator-demodulator 48, transceiver circuit 42 and interface circuit 46 operate under the control of processor 50, as described in more detail below.

An aspect of the operation of the base station 40 of being controlled by processor 50 is that user assignment is assigned to available channel to specific channel and by specific power level.The power of bringing up to a specific user's signal conventionally can improve can be provided for this user service (for example; by improving available data rate); but may worsen the service (for example,, by improving the level of the interference that they detect) that can be provided for other user.

Here we suppose that we have K community in system, with J subband in addition, we suppose that each subband consists of the subcarrier of fixed number.And the time of hypothesis is divided into time slot, and the transmission in each community is synchronous, thereby do not have the interference in community.Two general values are especially relevant to the Inter-Cell Interference Coordination mechanism of system for based on LTE.

First value is the concept of utilance, and utilance quantizes the satisfaction of related entity conventionally.Suppose that U is the overall utilization function of system, is provided by following formula:

U＝Σ _kU _k (1)

This formula represents the summation of all utilance functions on all communities, wherein U _kthe utilance function of community k, by the utilance U of all users for community k _{k, i}summation provide, i.e. U _k=∑ _iu _{k, i}.This idea is find a kind of mode (or various ways) to improve or preferably maximize overall utilization function U.

Second value is delivering power.Here, in the context of orthogonal frequency-division access (OFDMA) system such as LTE, delivering power is contemplated to relevant with frequency.Suppose P _{k, j}the power distributing in the subband j of Shi community k, and the maximum power that community k can have is P _k, i.e. ∑ _jp _{k, j}≤ P _k.The whole problem reduction of the interference coordination of minizone is P _{k, j}, how to be assigned to each k to improve or maximization U.

At A.L.Stolyar, H.Viswanathan, " Self-organizing Dynamic Fractional Frequency Reuse for Best-Effort Traffic Through Distributed Inter-cell Coordination (for the self-organizing dynamic part channeling of the optimal flux by coordinating between distributing cell) ", proc.ofIEEE Infocomm, in April2009, a kind of method based on gradient is suggested, wherein overall utilization by distributed way suboptimum be enhanced.The main idea of the method proposing is as follows:

Suppose for the utilance function U of community k _kwith respect to community m, distribute to the rate of change of the delivering power of subband j.The variation of the satisfaction that this value causes at the subband j place of community k corresponding to community m.For the object of discussing, this value is also referred to as D value for the sake of simplicity.Obviously, when k ≠ m (being that community m is contiguous community), P _{m, j}raising may be to U _khave a negative impact, because such raising is the additional interference producing from the subband j place of community m, vice versa.On the other hand, when k=m, the raising of the power at subband j place will strengthen the signal quality at this particular sub-band place, and the utilance to himself is had to active influence.

If note being considered in the Yi Ge community of same time, so D _j(m, k) is not of great use.But when it is exchanged between contiguous community, it allows contiguous community to know the level of the impact causing when specific power level is dispensed on each subband place in other community.By receiving the D from contiguous community _j(m, k), community k will add up to them for each subband subsequently,

D _j(k)＝∑ _mD _j(k，m)， (2)

(wherein the exchange of index m and k represent community k current be the fact of community of the contiguous m of each the contiguous community in contiguous community m), comprise the situation of k=m.

In other words, D _j(k) the total susceptibility of the utilance function causing corresponding to the disturbance due to himself delivering power at subband j place to all communities.

Work as D _j(k) during < 0, positive increased power has a negative impact the total satisfaction on all communities, and vice versa.The total idea proposing in prior art document discussed above is for community k, by selecting and D _j(k) subband that positive peak is associated improves power, and vice versa.

Suppose that δ P>0 is preset parameter, supposes P _k=Σ _jp _{k, j}for presently used gross power, and hypothesis for power limit.At n _pin each time slot in individual time slot, community k upgrades power as follows in turn:

1. set wherein j* is subband index, makes be minimum in all j, suppose D _j(k) < 0 and P _{k, j}>0.

2. if set j wherein ^*be subband index, make be maximum in all j, suppose D _j(k) >0.

3. if and max _jd _j(k) >0, sets and $P_{k,j^{*}}=P_{k,j^{*}}+\min (P_{k,j},\mathrm{\&delta;P}),$ Wherein with p _{k, j}>0 and those j in maximum and minimum.

In this illustrated embodiment, downlink power regulates algorithm to consider the service quality (QoS) that user experiences.

For the general utilance function of community k, be generally defined as the summation of logarithm of the mean bit rate of all users in the k of community.This utilance function derives from economics, and by the following true generation that excites, and this fact is and the user who enjoys high bit rate, compares for those, and fixing bit rate raising is more important for low bit rate.Another advantage of this function is that it is level and smooth and continuously differentiable divides, thereby simplifies the complexity while calculating utilance susceptibility.Although there is above advantage, such utilance function being not easy to provides a kind of QoS is considered to the mode in power adjustment mechanism.For example, consider, by three users of base station service, to there is respectively the bit rate of 1Mbps, 2Mbps and 3Mbps.If all three users only need the bit rate of 500kbps, from the angle of resource utilization, provide that to surpass necessary bit rate may not be efficient.Unnecessary high power produces unnecessary interference level, and then this interference has chain (knock-on) impact to contiguous community.In order to tie up good level of satisfaction,, by the higher power of needs, disturb thereby promote general background contiguous community.Vice versa: if its power is down to the level that just in time meets the requirement of user's bit rate in base station, to its neighbours' interference level, will reduce.Neighbours and then needs power is still less kept to call quality, thus the lower interference to origination base station sent.Therefore, origination base station and then needs power is still less kept to call quality.Till this process is continued until that the delivering power of ambient interferences and all base stations therefore is finally set as lower level.

The implication of above process is very important, because remove the idea of unnecessary power, provides a kind of " feedback " mechanism, and this mechanism finally contributes to due to the reduction of overall interference further to reduce for the fixing power requirement of QoS.The reduction of this power requirement is converted into for the energy of network and saves.

A mode considering QoS is to revise utilance function.Yet such method may make utilance function more complicated, thereby and make susceptibility calculation of complex.In this embodiment, we are by being called as the value H of " the happiness factor " _{k, i}whether user's expection is satisfied and is quantized, and this value is provided by following formula:

$H_{k,j}=\frac{{\stackrel{\&OverBar;}{R}}_{k,i}}{{\tilde{R}}_{k,j}}---\left(3\right)$

Wherein:

the mean bit rate that the user i in the k of community reaches, and

be corresponding bit rate requirement, this value can be proportional to assurance bit rate (GBR) (for example,, as 3GPP TS36.413, S1 Application Protocol (S1AP), discusses in issue 9, v9.5.1) or can be certain function of GBR.

Work as H _k,iduring >1, user experiences the bit rate that surpasses expection.Work as H _{k, i}during < 1, otherwise.Suppose for n of the happiness of community k weighted value constantly,

$H_k^{\left(n\right)}=\frac{1}{N_k}\underset{i=1}{\overset{N_k}{\mathrm{\&Sigma;}}}{w}_{k,i}{H}_{k,i}^n---\left(4\right)$

Wherein

N _kthe number of the user in the k of community, and

W _{k, i}it is the cell-specific weight for the user i in the k of community.

This weight can be used to apply and lay particular stress on a little between the user in community, and follows constraints $\underset{i=1}{\overset{N_k}{\mathrm{\&Sigma;}}}{w}_{k,i}=N_k.$ As special circumstances, when $w_{k,1}=w_{k,2}=...=w_{k,N_K}=1$ Time, ${\stackrel{\&OverBar;}{H}}_k=H_k^{\left(1\right)}$ Be reduced to simple arithmetic average.

Note mean that the average happiness for community k meets expection.Yet it also means that some users are below expection, and some users are more than expection.Although for quantization performance, is in general on average useful, more accurate method be for mean value provide guard surplus so that

$H_k={\stackrel{\&OverBar;}{H}}_k-{\mathrm{\&lambda;}}_k{\hat{H}}_k,---\left(5\right)$

H wherein _kbe called as " truly " happiness, and λ _kit is the scale factor of controlling " guarding " level.Value be the weighted standard deviation of the happiness in the k of community, and provided by following formula:

${\hat{H}}_k=\sqrt{H_k^{\left(2\right)}-{\left(H_k^{\left(1\right)}\right)}^2}---\left(6\right)$

The conservative effect that the happiness factor is offset to improve for power adjustments is illustrated in Fig. 3.Fig. 3 shows H _kprobability density function.Utilize power management operation to make the true happiness of cartographic represenation of area of region A in Fig. 3 is lower than the probability of the l of unit.By the happiness factor is offset true happiness is reduced to the area of region B lower than the probability of unit 1 from the area of region A.

Fig. 4 illustrates a kind of more general mode that improves conservative, by by H _kbe defined as H _{k, j}, x percentile (percentile).Under this definition, only have the X% of happiness will the unit of dropping on below 1, as shown in Figure 4.

Fig. 5 shows the process that downlink power is set, and the true happiness factor is taken into account.This process is periodically repeated.

In step 70, j* is selected for subband index, so that at D _j(k) < 0 and P _{k, j}in the situation of >0, minimum in all j.Thereby the power that this step is selected for it reduces and will produce the subband of best effect.

This process proceeds to step 72 subsequently, and wherein, in subband index J*, power is lowered.Particularly, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby, $P_{k,j_{*}}=\max (P_{k,j_{*}}-\mathrm{\&delta;P},0).$

After completing steps 72, this process proceeds to step 74.In step 74, determine for community P _ktotal delivering power whether be less than allowed maximum gross power .

If the total delivering power for community is less than allowed maximum gross power, power can be enhanced in a subband in a plurality of subbands, and this subband is selected in step 76.Thereby the power that step 76 is selected for it improves and will produce the subband of maximum advantageous effects.That is to say subband j ^*selected, so that maximum in all j, D wherein _j(k) >0.

The happiness of community is used to determine in fact whether will improve the power in described subband subsequently.Particularly, process proceeds to step 78, and wherein whether test cell is happy.By testing true happiness, whether, whether the unit of being less than 1 (that is have H for this _k< 1) determine.If this condition is satisfied, determine community happiness not, and process proceeds to step 80, wherein in step 76 in selected subband, power is enhanced.Particularly, power is from its currency be enhanced and increase progressively value δ P, or the gross power that is enhanced Bu Shi community is brought up to the gross power that maximum is allowed to the maximal increment that can be employed in above situation (if increment is below less).In other words, step 80 is set $P_{k,j^{*}}=P_{k,j^{*}}+\min (\mathrm{\&delta;P},{\tilde{P}}_k-P_k)$

If determine community enough happiness, i.e. H in step 78 _k>=1, power is lowered, to save energy and to raise the efficiency.Particularly, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby step 82 is set $P_{{k,j}^{*}}=\max (P_{{k,j}^{*}}-\mathrm{\&delta;P},0).$

If determine and used for the maximum gross power of community in step 74, power can only be enhanced in the situation that it is also lowered in another subband in a subband.Therefore, if determine that in step 74 inequality is not true, process proceeds to step 84, and wherein subband is selected.Thereby, based on d in all h _j(k) maximum, and for meeting P _{k, j}the different value D of the j of >0 _j(k) minimum value, subband j ^*be selected as optimum and carry out power raising, and subband j* is selected as optimum and carries out under powered.

The in the situation that of having selected current optimum carry out under powered subband in step 84, process proceeds to step 86, and wherein power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby step 86 is set

Then in step 86, by redistributing the power of removing in a subband from a plurality of subbands, judge that whether the power in the subband improving in a plurality of subbands is favourable.Particularly, in step 88, whether test cell is happy.By testing true happiness, whether, whether the unit of being less than 1 (that is have H for this _k< 1) determine.If this condition is satisfied, determine community happiness not, and process proceeds to step 90, wherein in step 86 in selected subband (wherein power improves the subband with maximum advantageous effects), power is enhanced.Particularly, power is enhanced the amount that the power in subband j* is lowered in step 86.Thereby power is from its currency be enhanced and increase progressively value δ P, or the power before being enhanced in subband j* (if the latter's amount is less).In other words, step 90 is set $P_{k,j^{*}}=P_{k,j^{*}}+\min (\mathrm{\&delta;P},P_{k,j_{*}}).$

If determine community enough happiness, i.e. H in step 88 _k>=1, power is lowered, to save energy and to raise the efficiency.Particularly, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby step 92 is set $P_{k,j^{*}}=\max (P_{k,j^{*}}-\mathrm{\&delta;P},0).$

Thereby process is tested the whether unit of being less than 1 of true happiness in step 78 and 88, and step 82 and 92 can arrange the performance number lower than the value that will arrange in other situation in the situation that of the true happiness unit of being more than or equal to 1.

Fig. 6 illustrates as proportionality factors lambda _kselect the impact of different values on resulting utilance.Particularly, Fig. 6 to show target transport block size (TBS) be the 70 per delivery time intervals of byte (TTI), for λ _k=0.01 (lines 100 in Fig. 6) and λ _kthe average utilization of each community of the function as the time of=1.20 (lines 102 in Fig. 6) (in this illustrative embodiment, this average utilization as defined above, is the summation of logarithm of all users' mean bit rate).For comparison purposes, for not having the result (lines 104 in Fig. 6) of the situation of qos requirement to be also included.At λ _k=0.01 place, can find out reasonably tracking target (lines 106 in Fig. 6) of average utilization.Along with λ _krise to 1.20, it is more conservative that system becomes, and causes the raising of the average utilization of expection, thereby and cause reached bit rate to drop on the minimizing of the number of the user below desired bit rate.In the situation that not applying QoS limit value, system will be used power as much as possible, and utilance is correspondingly higher.

Fig. 7 illustrates as proportionality factors lambda _kselect the impact of different value on the average power of each community.Particularly, Fig. 7 shows for λ _k=0.01 (lines 110 in Fig. 7) and λ _kthe average power of each community of the function as the time of=1.20 (lines 112 in Fig. 7).For comparison purposes, for not having the result (lines 114 in Fig. 7) of the situation of qos requirement to be also included.Thereby although system is issued to higher utilance in the situation of not considering QoS limit value, delivering power is also higher, because each eNB transmits with its maximum power 10mW, as shown in Figure 7.On the other hand, when the happiness factor is employed when considering qos requirement, can realizes significant power and save.At λ _kin=0.01 situation, utilance is compared with the situation that there is no QoS limit value and is reduced approximately 8%, and delivering power reduces and surpasses 80%.The correspondingly less reduction of the less reduction of utilance and subsidiary delivering power can realize by higher, more conservative scale factor is set.The reduction of this delivering power may not be attributable simply to the bit rate requirement of reduction, and weakening owing to overall interference.This demonstration can realize significant power and save.

As the above mentioned, proportionality factors lambda _kcontrol in such a way " conservative " of utilance, make higher scale factor value take higher delivering power and improve the overall utilance of system as cost.Thereby this parameter provides the degree of freedom of carrying out the utilance level of tuning system by the balance between utilance and power consumption.

For example, scale factor can be conditioned based on current delivering power.If current delivering power reaches its maximum, systematic function can not constantly improve.On the other hand, by reducing λ _k, the small size reduction of overall utilance may occur, but may occur may very significant power reduction.Thereby one of above algorithm is substituted is if average power higher than certain threshold value, make λ _kreduce step delta λ _kand if average power makes λ lower than another threshold value _kincrease step delta λ _k.

Work as proportionality factors lambda _kwhen higher, more power is used to improve the overall utilance in community, thereby improves the level for neighbours' dl interference.Thereby it is that the level of locating viewed dl interference based on community is carried out the λ in adaptive community that another of above algorithm substitutes _k.Interference value can utilize down link listen mode (DLM) or traverse measurement result, access (E-UTRA) carrier wave received signal strength indicator symbol of the UMTS terrestrial radio based on evolution (RSSI) and obtained.These measurement results can be based on time average RSSI value or the percentiles based on RSSI value.Reduce λ _kthe suggestion of value is for example transmitted to neighbours by the private message in X2 interface subsequently.

As discussed above, value D _j(k, m) describes owing to changing the level of the utilance impact causing from the power of neighbours m at subband j place.Then the total impact, changing for all neighbours' of subband j utilance is provided by following formula:

$D_j\left(k\right)=\underset{m}{\mathrm{\&Sigma;}}{D}_j(k,m)---\left(7\right)$

The calculating D proposing in Stolyar _j(k) method thereby relate to utilance function with respect to the calculating of the rate of change of power.This value is added up to subsequently, and as shown in equation (2), and result is distributed to neighbours.But, do not have the standard interface (particularly, such value is not supported by standard x 2 interfaces) that allows such value to transmit between base station, and therefore the transmission of this result needs the exclusive interface between base station.Therefore due to exclusive interface being provided, may be, inconvenient or impossible, and if standard x 2 interfaces can be more efficient so that the algorithm of the base station use that can only be connected by X2 interface to be provided with using standard x 2 interfaces, so D _j(k) place of equation of value is used.

According to TS36.423, X2 application protocol (X2AP), V8.3.0,3GPP, 2008.0, arrowband Tx power (RNTP) information element (IE) is included in load information X2 message relatively.For each Resource Block, RNTP IE inform contiguous community at the transmission cell power at this Resource Block place higher than (1) or lower than (0) certain threshold value (RNTP threshold value).In order to use X2-interface, we need to by X2 can with item write out D _jthe formula of (k, m).For D _jthe value through rewriting of (k, m) subsequently can be exchanged between base station.

As an example of such rewriting, establish:

$D_j(k,m)=\left\{\begin{array}{cc}-\underset{i}{\mathrm{\&Sigma;}}\frac{G_i^{\left(m\right)}}{G_i^{\left(k\right)}}{\mathrm{\&rho;}}_{j,m}& m\&NotEqual;k\\ 0& m=k\end{array}\right.---\left(8\right)$

Wherein the path gain between mobile device (k serves by community) and neighbours m, and ρ _{j, m}the RNTP for subband j.

Therefore, mobile device can be measured from contiguous community, to obtain this information, and can report back serving BS.Serving BS can calculate based on equation (8) subsequently.

Path gain can pass through measuring reference signals reference power (RSRP) (at TS36.214, physical layer; Measure, V9.2.0,3GPP, is described in 2010) and at mobile device place, be obtained from the corresponding delivering power of neighbours m via neighbours' broadcast channel.

More accurately, being located at the path gain that time t place is sampled is:

$g_i^{\left(m\right)}\left(t\right)=\frac{\mathrm{RSR}{P}_m}{P_m^{\mathrm{ref}}}---\left(9\right)$

RSRP wherein _m(t) be the RSRP from community m being sampled at time t place, and it is the reference signal power from neighbours' broadcast channel.

As an alternative, can be defined as because the ratio of RSRP value also will provide the relative effect of neighbor base stations m with respect to serving BS.

The RSRP measurement result that attention (utilizes near the down link watch-dog (DLM) base station) at mobile device or at base station place and obtains may fluctuate due to channel fading, blanking etc.What conventionally adopt a lot of samples is on average on average more representative with what extract path gain for a long time.Thereby, can be used as index is on average obtained: $G_i^{\left(m\right)}\left(t\right)=(1-a)G_i^{\left(m\right)}(t-1)+a{g}_i^{\left(m\right)}\left(t\right),$ Or as block, be on average obtained more simply: $G_i^{\left(m\right)}\left(t\right)=\frac{1}{N}\underset{I=1}{\overset{N}{\mathrm{\&Sigma;}}}{g}_i^{\left(m\right)}(t-i).$

Alternately, it can be sample $\{g_i^{\left(m\right)}\left(t\right),t=t-1,t-2,...,t-N\}$ Percent x.

Value ρ _{j, m}it can be the reasonable aggregate value for the RNTP value of each Resource Block in subband.A kind of scheme of simply resolving is

${\mathrm{\&rho;}}_{j,m}=\frac{1}{Q}\underset{q=1}{\overset{Q}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}}_{j,m}^{\left(q\right)}---\left(10\right)$

Wherein Q is the number of the Resource Block of each subband, and the RNTP for the Resource Block q in the subband j from neighbours m.The another kind of mode that the value of each Resource Block is added up in subband is for example to get , in maximum.

Fig. 8 is the overview flow chart of the process of periodicity repetition.Note value D _j(k, m) is no longer specifically shown in the utilance of subband j Chu community k with respect to the susceptibility of the power from neighbours m.But it is illustrated in community k will be due to the total impact in the mobile device in the k of m Er community, community in the situation that subband j place transmits.This value is less negative, and its impact is larger, and therefore, community k is by the transmission of further avoiding it at relevant subband place.

Be important to note that the D in above equation (8) _jthe formula hypothesis mobile device of (k, m) has the ability of measuring broadcast channel, and obtains the direct measurement result of neighbours' delivering power and the RSRP of same neighbourhood.A kind of calculating D _jthe simplified way of (k, m) is provided by following formula:

$D_j(k,m)=\left\{\begin{array}{cc}-\frac{G^{\left(m\right)}}{G_k}{\mathrm{\&rho;}}_{j,m}& m\&NotEqual;k\\ 0& m=k\end{array}\right.---\left(11\right)$

G wherein ^(m)the path gain residing between the corresponding transmitter of locating for the base station of the DLMYu community m base station of community k near, and G _kit is certain positive constant.

In formula in equation (11), D _jthe larger value of (k, m) is more approaching zero value.If the path gain the neighbours at j place is larger, and these neighbours are with as ρ _{j, m}indicated higher power transmits, and ratio has larger value, and the negative sign before this ratio to make this value be less negative, and more away from zero.This value is less negative, and this subband k is more unfavorable to transmission.Thereby if serving BS will transmit at subband j place, susceptibility and risk are therefore higher.

If there is no the mobile device of report information can use, permission is calculated at serviced mobile device and the path gain between neighbours, serving BS still can rely on its down link watch-dog (DLM), wherein it detects the signal being sent by neighbor base stations in the down-link frequencies of system, to make the estimation to path gain (between its DLM and neighbours).In other words, for the object that path gain is estimated, DLM shows to obtain picture user.Certainly, this does not represent from mobile subscriber and obtains information, because in the diverse location of mobile subscriber in community.

Finally, if there is no DLM Information Availability, base station relies on the information that X2 interface provides, i.e. ρ by having to _{j, m}value, and so D _j(k, m) can be defined as:

$D_j(k,m)=\left\{\begin{array}{cc}-{\mathrm{\&rho;}}_{j,m}& m\&NotEqual;k\\ 0& m=k\end{array}\right.---\left(12\right)$

Thereby the X2 compatible version of this algorithm is rewritten D according to a kind of like this mode _j(k), so that its utilization can be caught data available in the X2 interface between two eNB.As example, D _j(k) can be redefined by the mode shown in above equation (8), (11) or (12).

As the result redefining, some modifications of primal algorithm are made so that this algorithm is more stable and robust.

Thereby in Fig. 8, in step 120, j* is selected for subband index, so that at D _j(k) < 0 and P _{k, j}in the situation of >0, minimum in all j.Thereby the power that this step is selected for it improves the subband that estate performance is produced to minimum or minimum advantageous effects, and the power of this subband after may be lowered.

In step 122, subband index j ^*selected, object is to select for its power to improve the subband that estate performance is produced to maximum or advantageous effects.As shown in equation (7) and (11), at optimal sub-band place, D _j(k) by having, be zero value, and may have a plurality of subbands that meet this standard.The possibility being enhanced in a subband only for fear of power, wherein power may after the subband that is enhanced from thering is D _j(k) in one group of subband of=0, selected at random.In this manner, may have more subband can there is non-zero power, and subband utilance improve.

In step 124, whether test has .Due to subband index j* selected so that minimum in all j, thus step 124 test in fact whether exist have for any subband of negative value.In other words, whether step 124 test exists the power for it to improve and will have any subband of unfavorable effect.

In step 124, whether all right test cell is happy.This is by the true happiness H of test _kwhether be greater than service quality (QoS) and require η _kwith hysteresis factors ζ _kproduct determine, i.e. whether test has a H _k> η _kζ _k.The value of qos requirement is the value of being set to 1 for example.

If found in step 124 and in test cell, whether also find H during happiness _k> η _kζ _k, process proceeds to step 126, and wherein power is lowered in subband index j*.Particularly, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.

After completing steps 126, if or find not exist the power for it to improve and will there is subband and/or the true happiness H of unfavorable effect in step 124 _kbe not more than service quality (QoS) and require η _kwith hysteresis factors ζ _kproduct, process proceeds to step 128.

In step 128, determine for community P _ktotal delivering power whether be less than allowed maximum gross power and determine that whether in step 122, in selected subband, improve power favourable simultaneously, whether have

If these conditions are all satisfied, process proceeds to step 130, and wherein whether test cell is happy.Whether this is less than qos requirement η by testing true happiness _k(that is, H _k< η _k) determine.The in the situation that of qos requirement value of being set to 1, by testing true happiness, whether, whether the unit of being less than 1 (that is have H for this _k<1) determine.If this condition is satisfied, determine community happiness not, and process proceeds to step 132, wherein power is enhanced in selected subband in step 122.Particularly, power is from its currency be enhanced and increase progressively value δ P, or the gross power that is enhanced Bu Shi community is brought up to the gross power that maximum is allowed to the maximal increment that can be employed in above situation (if increment is below less).In other words, step 132 is set $P_{k,j^{*}}{P}_{k,j^{*}}+\min (\mathrm{\&delta;P},{\tilde{P}}_k-P_k)$

If determine community enough happiness, i.e. H in step 130 _k>=η _k(or the in the situation that of qos requirement value of being set to 1, H _k>=1), process proceeds to step 133, wherein determines true happiness H _kwhether be greater than qos requirement η _kwith hysteresis factors ζ _kproduct, judged whether H _k> η _kζ _k.The in the situation that of qos requirement value of being set to 1, be actually the true happiness H of judgement _kwhether be greater than hysteresis factors ζ _k, judged whether H _k> ζ _k.

If determine that in step 133 true happiness is more than sufficient, i.e. H _k> η _kζ _k(or, the in the situation that of qos requirement value of being set to 1, H _k> ζ _k), power is lowered to save power and raises the efficiency.Particularly, power is from its currency reduce the value δ that successively decreases _pbut power can not be lowered to below zero certainly.Thereby,

If determine and used for the maximum gross power of community in step 128, or raising power is not favourable for optimal sub-band, and process proceeds to step 136, wherein determines for community P _ktotal delivering power whether equal allowed maximum gross power and judge that whether in step 122, in selected subband, improve power favourable simultaneously, whether have if these conditions are not satisfied, algorithm stops and waiting until next time always and carry out and start.But, if these conditions are satisfied, show to be still worth doing further power adjustments.

When the total delivering power for community has reached the gross power that maximum is allowed to,, regulate and need the subband that wherein power can be lowered.Thereby, in step 138, test and whether have even whether have the subband of minimum in power also can advantageously be enhanced.If this condition is satisfied, process proceeds to step 140.

In step 140, new subband is from having D _j(k) in one group of subband of=0, be selected randomly, and process proceeds to step 142 subsequently.

Or, if found in step 138, have the subband of minimum have or more specifically have be that power can advantageously be lowered in this subband, process is directly to step 142.

In step 142, in the subband that power is found in step 138, be lowered to have minimum negative value, or in selected subband, be lowered in step 140.Have, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby step 142 is set

Then judge by redistribute whether the power that the power of removing in a subband from subband improves in a subband in subband is favourable in step 142.Particularly, in step 144, whether test cell is happy.By testing true happiness, whether whether the unit of being less than 1 (has H for this _k< 1) determine.If this condition is satisfied, determine community happiness not, and process proceeds to step 146, wherein in step 122 in selected subband (wherein power improves the subband with maximum advantageous effects), power is enhanced.Particularly, power is enhanced the amount that the power in subband j* is lowered in step 142.Thereby power is from its currency be enhanced and increase progressively value δ P, or the power (if the latter is less) before being enhanced in subband j*.In other words, step 146 is set $P_{k,j^{*}}=P_{k,j^{*}}+\min (\mathrm{\&delta;P},P_{k,j_{*}}).$

If determine community enough happiness, i.e. H in step 144 _k>=η _k(or the in the situation that of qos requirement value of being set to 1, H _k>=1), process proceeds to step 147, wherein judges true happiness H _kwhether be greater than qos requirement η _kwith hysteresis factors ζ _kproduct, judged whether H _k> η _kζ _k.The in the situation that of qos requirement value of being set to 1, be actually the true happiness H of judgement _kwhether be greater than hysteresis factors ζ _k, judged whether H _k> ζ _k.

If in step 147, definite true happiness stamp stamp is had a surplus, i.e. H _k> η _kζ _k(or, the in the situation that of qos requirement value of being set to 1, H _k> ζ _k), power is lowered to save power and raises the efficiency.Particularly, power is from its currency reduce the value δ P that successively decreases, but power can not be lowered to below zero certainly.Thereby,

Thereby step 140,142 and 146 combined effect are the power reducing in a good subband, and improve the power in another good subband.This has created chance for system and between subband, has redistributed power and randomization and the diversified power that is assigned to subband to be avoided falling into local maximum.

Whether a kind of replacement of the algorithm shown in Fig. 8 is by the total delivering power about for community, whether to be less than the gross power that the maximum of community is allowed to (that is, to have ) single judgement replace Decision Block 128 and 136.In this manner, power improves not necessarily subband by the time not to be disturbed completely, and base station can be improved at the power with the subband place of least interference.

Providing the slightly extensive another kind of alternative of the algorithm in Fig. 8 is to replace the selection in step 120 by the random selection to j*.That is to say, j* can be from set omega _kmiddle quilt is selected at random, wherein Ω _k=(1), (2) ..., (M _k), 1≤M _k≤ J and (j) be and D _j(k) i the index that minimum value is corresponding, i.e. D ₍₁₎(k)≤D ₍₂₎(k)≤...≤D _(J)(k).Work as M _k=1 o'clock, this was reduced to the original steps 120 in Fig. 8.This extensive object be randomization and variation to the selection of j* further to improve and optimizate result.

Fig. 9 shows and is the 90 per delivery time intervals of byte (TTI), selects for λ when target transport block size (TBS) _k=0.01 (lines 160 in Fig. 9) and λ _kduring the different value of=1.20 (lines 162 in Fig. 9) as the average utilization of each community of the function of time.For comparison purposes, for not having the result (lines 164 in Fig. 9) of the situation of qos requirement to be also included.Similar with above Fig. 6, feasible in the situation that, the average utilization of each community converges to target (lines 166 in Fig. 9).But convergence time is slightly longer, especially for λ _k=0.01 situation.

Figure 10 illustrates for proportionality factors lambda _kselect the impact of different value on the average power of each community.Particularly, Figure 10 shows for λ _k=0.01 (lines 170 in Figure 10) and λ _kthe average power of each community of the function as the time of=1.20 (lines 172 in Figure 10).For comparison purposes, for not having the result (lines 174 in Figure 10) of the situation of qos requirement to be also included.Thereby for rational bit rate target, power efficiency is also very high.

As mentioned above, scale factor can be conditioned based on current delivering power or the level based on viewed dl interference.

Can find out the energy efficiency that the total QoS limit value supported in mobile device and the gap between cell capacity limit community.In other words, when adding up to QoS limit value higher than cell capacity, all power will be used, and can not have power to save.But by QoS limit value is down to slightly lower than cell capacity, power is saved and is started to become possibility.

Another proposal that realizes energy saving is dynamically to carry out to reduce adaptively QoS limit value by observing cell throughout.

In order to obtain susceptibility D as above _j(k, m), we need the derivative of calculation plot utilance to the delivering power in subband j Chu community m.Conventionally, community utilance is relevant with the spectrum efficiency of subband.For example, establish

$\begin{array}{c}{D}_j(m,k)=\frac{{\&PartialD;U}_k}{\&PartialD;P_{m,j}}\\ \&ap;\underset{i\&Element;{\mathrm{\&Omega;}}_k}{\mathrm{\&Sigma;}}f\left(\frac{\&PartialD;\mathrm{\&omega;}\left({\mathrm{\&gamma;}}_{i,j}^{\left(k\right)}\right)}{{\&PartialD;P}_{m,j}}\right)\\ =\underset{i\&Element;{\mathrm{\&Omega;}}_k}{\mathrm{\&Sigma;}}f(\frac{\&PartialD;\mathrm{\&omega;}\left({\mathrm{\&gamma;}}_{i,j}^{\left(k\right)}\right)}{\&PartialD;{\mathrm{\&gamma;}}_{i,j}^{\left(k\right)}}\&CenterDot;\frac{\&PartialD;r_{i,j}^{\left(k\right)}}{\&PartialD;P_{m,j}})\end{array}---\left(13\right)$

Ω wherein _kthe user's index set in the k of community, be signal and interference and the noise ratio (SINR) of the user i in subband j Chu community k, and ω is conduct the spectrum efficiency of function.Item f () is to define D _jthe pre-defined function of (k, m), and be therefore that base station is known.Value relatively direct, because p _{m, j}known function.Subject matter is value this value depends on the essence of ω, and is not explicitly known in practice.As a rule, often suppose

$\mathrm{\&omega;}\left(x\right)={\log }_2(1+\frac{x}{\widehat{\mathrm{\&Gamma;}}})---\left(14\right)$

Wherein the fixed value that is often called as " capacity-gap " constant, how many because it determines with actual channel capacity difference.For the sake of simplicity, at prior art document R.Kwan, C.Leung, " A Survey of Scheduling and Interference Mitigation in LTE (investigation that the scheduling in LTE and interference suppress) ", volume 2010, supposes in article ID273486 and suppose in other prior art document (wherein, ε _bbit error rate), described other prior art document is A.J.Goldsmith for example, S-G Chua, " Variable-Rate Variable-Power MQAM for Fading Channels (for the variable bit rate variable power MQAM of fading channel) ", IEEE trans.on Comm.V01.45, no.10, in October, 1997; G.Piro, N.Baldo.M.Miozzo, " An LTE module for the ns-3network simulator (for the LTE module of ns-3 network simulator) ", in Proc.of Wns32011 (in conjunction with SimuTOOLS2011), in March, 2011, Barcelona (Spain); And H.Seo, B.G.Lee. " A proportional-fair power allocation scheme for fair and efficientmultiuser OFDM systems (for the equitable proportion allocative decision of fair and efficient multi-user OFDM system) ", in Proc.of IEEE GLOBECOM, in December, 2004, Dallas (USA).

It is important to note that the analysis relation described in above equation (14) is theoretic, because actual SINR is unknown for base station.According to 3GPP TS36.213, physical layer procedures, issue 9, v9.3.0, mobile device is measured downlink channel quality with the form of SINR, and to be called as the form of the index of CQI (CQI), such value is encapsulated.CQI is that reception base station is available.

In addition, equation (14) has defined a kind of static relation, even if this static relation is well approximate in specific environment, but may be not too accurate in another environment.In practice, the implementation that spectrum efficiency and the relational dependence of channel quality may be concrete due to the supplier of receiver architecture and different.Thereby, a kind of make ω with the mode of relevant more robust will be very useful.

According to 3GPP TS36.213, physical layer procedures,, between the CQI reporting of spectrum efficiency and mobile device, there is definite relation in issue 9, v9.3.0.In other words, once CQI is known, base station just can obtain reporting corresponding spectrum efficiency with each CQI by look-up table.Although the clear and definite analytic relationship between spectrum efficiency and CQI is not presented, we propose very reasonably spectrum efficiency to be approximately the power function of CQI:

ω＝aq ^b(15)

Wherein q is CQI, a=0.077 and b=1.586.

Although the relation between spectrum efficiency and CQI is fixed, the mode that CQI is measured is not by standardization, but it most possibly depends on measured SINR.The mode that SINR is measured is vendor-specific, and depends on a plurality of factors, comprises the related implementation of receiver algorithm, the accuracy of estimation etc.But in practice, it is to be substantially linear as the function of the SINR of YidBWei unit that CQI is designed to make it according to a kind of like this mode.Each point at given CQI value place in such linear relationship has similar distance to its neighbours.Although may there is the variation of the concrete implementation of supplier, such relation can not depart from much each other, because in the situation that the standardized algorithm of given sender side only has the rational method of a limited number of receiver designing.A good example of the relation between CQI and SINR can be at prior art document C.Mehlf ü hrer, M.Wrulich, J.C.Ikuno, D.Bosanska, M.Rupp, " Simulating the Long Term Evolution Physical Layer (simulation Long Term Evolution physical layer) ", the 17th European signal processed the Proc. of meeting (EUSIPCO), found in 2009.

Experience to such relation is estimated as:

q＝cγ _dB+d (16)

γ wherein _dBsINR (thereby the γ representing with decibel _dB=10.log ₁₀γ, wherein γ is SINR), c=0.5 and d=4.4.Thereby, as the spectrum efficiency of the function of SINR, by following formula, provided

ω＝a(c′log ₁₀(γ)+d) ^b (17)

C '=10c wherein.

As mentioned before, CQI is not standardized about SINR curve, and between supplier, may have the difference of implementation slightly.However, we can compensate to obtain to such difference by equation (17) is introduced to offset Δ d

ω＝a(c′log ₁₀(γ)+d+Δd) ^b (18)

Equation (18) itself is a kind of simple form of differential continuously, and its derivative is relatively simple.

Figure 11 shows the adjusting to the relation between spectrum efficiency ω and SINR by CQI offset Δ d.Particularly, introduce positive offset Δ d CQI is fastened and moved as the indicated relation of lines 182 from the indicated pass of lines 180 with respect to the relation of SINR, thereby and mean that higher CQI value is obtained for given measured SINR value.This point that also means that spectrum efficiency is higher with respect to curve 184 tops of CQI is selected, thereby selected for the higher value of spectrum efficiency ω.Note, from calculating D _jthe angle of (m, k), offset Δ d not necessarily needs to be quantized because main purpose be obtain the analysis of spectrum efficiency ω approximate so that can be calculated about the derivative of SINR.

Figure 12 shows the effect of utilizing offset Δ d to carry out CQI adjusting.As example, Figure 12 (a) shows the effect of introducing offset Δ d=-2, be that original CQI lines 190 are (based at prior art document C.Mehlf ü hrer, M.Wrulich, J.C.Ikuno, D.Bosanska, M.Rupp, " Simulating the Long Term Evolution Physical Layer (simulation Long Term Evolution physical layer) ", the 17th European signal is processed the Proc. of meeting (EUSIPCO), resulting result in 2009) be displaced to vertically downward lines 192.In other words for any given SINR value, lower CQI value is obtained.

Then, Figure 12 (b) shows by the skew of CQI value, (the Vienna University of Technology of prior art document Zhong You Tech Uni Wien at Mehlf ü hrer etc., VUT) the original signal spectrum efficiency curve of making (lines 194 in Figure 12 (b)) is offset to produce the more low value for spectrum efficiency, and resulting curve (lines 196 in Figure 12 (b)) with based on prior art G.Piro, N.Baldo.M.Miozzo, " An LTE module for the ns-3network simulator (for the LTE module of ns-3 network simulator) ", in Proc.of Wns32011 (in conjunction with SimuTOOLS2011), in March, 2011, hypothesis in Barcelona (Spain) is by telecommunication technology center, Catalonia (Centre Tecnol ò gic de Telecommunications de Catalunya, CTTC) the spectrum efficiency curve of independently making (lines 198 in Figure 12 (b)) coincide.

A kind of mode of determining the value of Δ d is to feed back by hybrid ARQ.If the ratio of the number of negative response (NACK) message and transmission (comprising transmission again) sum is greater than certain threshold value on the specific time period, Δ d is subtracted 1.On the other hand, if this ratio on the specific time period lower than specific threshold, Δ d is added 1.

Thereby, if there is a high proportion of NACK message, mean that channel quality is lower than expecting before, so deviant successively decreased, mean that the value of derived CQI is lowered.The value of lower CQI refers to more " guarding ", and is relatively not easy to make mistakes.

Thereby this provides a kind of mode to obtain empirical formula between spectrum efficiency and SINR, simple relation analytically.This is very important, because CQI can be used for the unique information of base station according to standard.Then, such relation can be used to obtain the sensitivity function for above-mentioned power management mechanism.This total relation provides a kind of mode will himself to be adapted for certain real bottom relation to regulate by simple parameter.

Then, Hybriad ARQ can be used to regulate parameter so that estimated relation is mated real bottom relation better.As substituting that HARQ is fed back, can use the difference between the average block error rate and corresponding desired value, i.e. the average BLER-target of X=BLER.If X is greater than 0 on time period T (or or even be greater than better little positive threshold value), side-play amount is lowered Yi Ge unit.On the other hand, if X on time period T lower than 0 (or little negative threshold value), side-play amount is increased Yi Ge unit.

Notice that above method needs the special purpose interface between two nodes, so that node can transmit the needed information of performance number that arranges as discussed above.Figure 13 illustrates a kind of may layout, millimicro microbedding 220 comprise a plurality of HeNB222a, 222b ..., 222k, and Macro 224 comprise a plurality of eNB226a, 226b ..., 226k.In millimicro microbedding 220, can and communicate with one another via X2 interface by all HeNB of reasonable assumption.This especially may occur in the environment that is called as " enterprise " environment, and wherein all HeNB may be from same provider.In Macro 224, all eNB also expect and can and communicate with one another via X2 interface.But, always may not there is the X2 interface between millimicro microbedding 220 and Macro 224.

Another related fields of layout in the network of the type shown in Figure 13 are frequency spectrum overlapping modes between Macro and millimicro microbedding.

Figure 14 shows three kinds of different possibilities, schematically illustrates available band and how between Macro and millimicro microbedding, to be divided.

The first possibility (situation A) is that subband 230 is assigned to Macro and subband 232 is assigned to millimicro microbedding, therefore between this is two-layer, does not have frequency overlapping.Thereby, need between this is two-layer, not carry out to disturb and suppress, because related frequency band is not common channel.

The second possibility (case B) is that subband 234 is assigned to Macro and subband 236 is assigned to millimicro microbedding, makes the frequency band of millimicro microbedding completely by the frequency band superimposition of Macro.Thereby, although the not overlapping region of Macro is not affected, on the impact of overlapping region, may be very large.In this case, interference management becomes very useful.

Possibility between (situation C) is that subband 238 is assigned to Macro and subband 240 is assigned to millimicro microbedding, and exists part overlapping between Macro and millimicro microbedding.The scheduler of the base station in each layer is contemplated to automatic selection subband to avoid the interference of minizone, and the relation between this subband in two-layer can utilize can by the evolved UMTS terrestrial radio access (E-UTRA) separately of exchanged they of X2 interface definitely radio channel number (EARFCN) and bandwidth mapped.But, thereby by comprising that power that power management is lower is assigned to the subband of larger interference and vice versa, still can expect higher performance.

Be given in below in the situation of the first method shown in above-mentioned Fig. 5 and in the situation that the second method shown in above-mentioned Fig. 8 allows the possible deployment scenario of the communication between each base station, in the first method, power management relies on according to calculated happiness value the setting of performance number, need the exclusive interface between base station, and in the second method, power management relies on the value that can transmit by the X2 interface of standard.

the base station in millimicro microbedding only

the method that needs the exclusive interface between base station

When the HeNB in geographic area belongs to same supplier or has the supplier of certain special arrangement, this method is suitable for.This may occur in " enterprise " environment, and wherein the unified shared wherein mobile subscriber in Femto cell is contemplated to the space of free-roaming.In this case, can define the exclusive message as " private message " on X2.

can use the method for X2 interface

In this case, HeNB can be used the X2 interface of standard for the object of power adaptation.The version based on X2 of standard need to not used private message in X2 interface.It is unrare that HeNB in service area belongs to same supplier (or different suppliers of shared specific arrangements).But if two base stations do not belong to same supplier, the HeNB that realizes described algorithm still may benefit from the standard message from its neighbours.Thereby this solution is more insensitive for the compatible problem between base station, as long as the X2 interface of their share standard.

the base station in Macro only

the method that needs the exclusive interface between base station

Contiguous macro base station can be from different suppliers, but this method is only to be applied in the situation from same supplier's base station.

can use the method for X2 interface

This solution does not need exclusive interface.Therefore, only the same in the above situation in millimicro microbedding with base station wherein, X2 interface can be used in the situation that not needing to use private message.

base station in millimicro microbedding and Macro, and between them, X2 interface can be used

the method that needs the exclusive interface between base station

Because a large amount of Femto cells are positioned at the expection under single grand region, the power possibility that each HeNB execution power adaptation is suitable for each Femto cell than the trial setting of Macro base station is convenient.If only have HeNB carrying out adaptation, need to be at place, Macro base station operate power set algorithm (at least not needing to be adapted for the power level in millimicro microbedding).Thereby, for Macro base station, there is not vendor compatibility problem.

can use the method for X2 interface

Equally, each HeNB can carry out power adaptation, and need to be at place, Macro base station operate power set algorithm.

Utilize the available X2 interface between Macro and millimicro microbedding, being suitable for reference to the described method of figure 8 shown in Fig. 8.Needed information can be embedded in relative arrowband Tx power (RNTP) information element in load information X2 message.For each Resource Block, RNTP IE inform contiguous community the transmission cell power at such Resource Block place more than certain threshold value (RNTP threshold value) (1) or below (0), this threshold value is another X2 parameter in 3GPP standard.The characteristic that depends on algorithm, traverse measurement the possibility of result is required to calculate the path gain with respect to base station.This can be by comparing mobile down link RSRP measurement result and the reference signal delivering power of the system information block (SIB) of the broadcast channel of community from contiguous to realize.If proprietary information is required, the private message in X2 interface can be used.

base station in millimicro microbedding and Macro, and between them, do not have X2 interface to use

When not having X2 interface available, owing to take by frequency cells, be the grand impact that is difficult to estimate interference of the specific vicinity on basis.Thereby the impact of presence of intercell interference must be by direct estimation.

For example, a kind of mode of estimating presence of intercell interference in the situation that not using X2 interface is that configuration and periodic move CQI measurement result in whole bandwidth.

First, these CQI measurement results from all mobile devices of encamping are collected in base station.These CQI measurement results are considered to be in subframe aspect as instantaneous, and therefore in order to estimate long-term presence of intercell interference, and base station is carried out average to these measurement results subsequently, and this can be such as based on exponential average or piece on average etc.

Because mobile device is in different geographical location, they are different with respect to the path gain of base station separately.Thereby, from the average CQI measurement result of each mobile device, with respect to its mean value separately, be normalized subsequently.

Serving BS is gathered the normalized average CQI measurement result of all mobile devices that belong to this base station subsequently, and the CQI for all mobile devices of each subband across whole bandwidth averages, and obtains the vector for the normalized average CQI measurement result of community wherein j is corresponding with the index of subband.

For lower than certain threshold value 's each entry j, value D _j(k) can be set as negative real number value.For example, D _j(k) can be: (a) fixing negative real number value; (b) from being uniformly distributed the value of picking out in U (a ,-b), wherein a and b are certain real positive value; (c) value (d) neighbours for example, with respect to the negative value (, being similar to above equation (8)) that is attached to the maximum summation of the path gain ratio between all mobile devices of serving BS; Or (e) only relate to down link listen mode (DLM) and do not rely on the version of (d) of traverse measurement result (being similar to above equation (11)).

Once realize this point, the algorithm shown in just can application drawing 8.

It should be noted that frequency selection may affect the accuracy of above estimation.In other words, wireless channel can have variation to a certain degree by expectability on frequency band.Thereby it is normally more accurate in the environment of Femto cell that expection is estimated, because time delay expansion is conventionally less.But this method of estimation is not limited to Femto cell environment.

For X2, be not present between Macro and millimicro microbedding but be present in the situation in millimicro microbedding, for it, D is set _j(k) set of subband index should be the union between the set obtaining from X2 interface and the set of using traverse measurement result.D _j(k) value can for example be used the path gain measurements given as above option (d) or (e) or their variant to obtain.

" the happiness factor " was more than discussed and can be defined as mean bit rate that user reaches divided by bit rate requirement .If very high for the ability that this bit rate requires to process such requirement with respect to system, system will attempt using power as much as possible to meet this requirement.

Figure 15 shows average utilization function and average power how (that is, the bit rate with each mobile device requires ) change." capacity " C when system _{k, sys}when requiring, the income of using whole power is very little, because the performance of itself is limited to the inherent limitations of system.Here, term " capacity " is broadly defined as in the situation that the maximum performance that the systems such as position of given bandwidth, mobile device can realize.On the other hand, when bit rate requires to approach also slightly lower than power system capacity, the leeway of saving for power starts to occur, and power efficiency starts to improve.When desired bit rate is down to power system capacity when following, system becomes and may be downgraded to cost with a small amount of bit rate and reduce delivering power.But due to the reduction of presence of intercell interference, the reduction that reduces the bit rate causing due to power can compensate by improving SINR again.In addition the impact that, bit rate reduces is also further absorbed with respect to the logarithmic relationship of bit rate by utilance.

Figure 16 shows average utilization function with average power between relation, the point on lines 250 represents to require (that is, for the bit rate of each mobile device ) the relation of different value.When desired bit rate reduces, the state of system starts to be moved to the left from an A.Therefore can find out bit rate requirement can be set to require to compare with peak power the value (for example,, in the region on lines 250 252) that realizes very large power saving and don't can cause the large punishment with regard to utilance reduces.

Conventionally, desired bit rate is controlled by the higher level of network.But, a kind of may be base station in the following manner bit rate requirement be set to lower value:

First, obtain happiness H _{k, i}n continuous sample.If there is H for N ' individual (wherein N '≤N) continuous sample at least _{k, i}< 1, and system transmits with whole power, be lowered step-length .This adjustment process is carried out very slowly, because the value of N and N ' and power setting algorithm frequency when called is compared larger.

This can be repeated until till being reduced to minimum permissible value, or till the reduction of desired bit rate means that utilance is reducing more quickly than average power.So, for example, may be process will repeat himself until 1) average community utilance with respect to the rate of change (being exactly derivative in simple terms) of average cell power higher than specific threshold or 2) average community utilance is lower than certain utilance threshold value, or 3) subset of user's utilance is higher than certain threshold value.As shown in Figure 16, derivative be on the occasion of.When bit rate requires to be down to certain level, the value of derivative starts to increase very rapidly.Therefore system requires be lowered and record this derivative along with bit rate, and when condition 1), 2) or 3) while being satisfied, the reduction of bit rate requirement stops.

Minimum tolerable bit rate can be set with mobile device each bit rate require to be associated.As an example, this minimum allowable value can be set to the predetermined score value that initial bit rate requires.This predetermined score value can be set as fixed value, for example, and 1/2 or 3/4.Alternately, this predetermined value can be set up based on discharge pattern.Thereby this predetermined value can be set to for the flow of some kind 1/2 and be set to 3/4 for the flow of some other kind.Minimum allowable value should be provided so that it prevents that user i from cut off by service all the time.

As mentioned above, user's happiness is defined as mean bit rate divided by the bit rate requirement for user.Thereby when bit rate requires to be lowered, user looks happier.When user's happiness improves, more seldom need system to improve power during power adaptation.This reduces average power.When average power reduction and bit rate requirement reduction, user's mean bit rate reduces.Conventionally, utilance function U is the function of bit rate.When mean user bit rate reduces, corresponding utilance reduces.But the bit rate that reduces unhappy user requires (supposing that bit rate requires still more than minimum tolerable bit rate) can reduce the total mean power of system.

Load calculation is the importance of LTE, and is correlated with in the context of access control, congestion control and load balance.The suitable quantification of cell load is required to judge whether community can the new carrying (bearer) of access.When cell load is very high, more multi radio carrying enters the call quality that can be prevented from keeping existing carrying.Once be admitted in system, cell load still may fluctuate due to the variations of the result as channel fading and mobility etc.Thereby system need to be tackled such fluctuation of load, and if necessary, some existing carryings may need to be dropped.

The simplest mode of calculation plot load is with respect to the sum of the Resource Block of bandwidth, to calculate the average number of the Resource Block using.A shortcoming of this method is that it often excessively estimates load, especially in the situation that there is optimum (best-effort) flow, thereby and may cause the insufficient utilization to resource.Definition for the more accurate mode of the cell load of LTE at R.Kwan, " the On Radio Admission Control for LTE Systems (about the radio access control of LTE system) " of R.Amott etc., proc.ofIEEE VTC-fall, is suggested in 2010.For calculation plot load, the ratio of the spectrum efficiency of each Resource Block of the number of the needed Resource Block of each carrying based on desired bit rate and user and being obtained.This value is the total normalization of the Resource Block in system bandwidth subsequently, and all movable carrying in system is sued for peace.

But this method hypothesis power spectrum density is constant across all bandwidth.When frequency selectivity power is controlled by use, this hypothesis is effective.But, in the situation that exist across the power of bandwidth, control (that is, each subband may present different power levels), such method may underestimation load, because power is distributed on all subbands unevenly, thereby reduce the availability of some subbands.

In order to overcome this problem, load can be defined as:

${\mathrm{\&rho;}}_k=\underset{i}{\mathrm{\&Sigma;}}\frac{{\tilde{R}}_{k,i}}{{\stackrel{\&OverBar;}{R}}_{k,i}}\frac{{\stackrel{\&OverBar;}{P}}_{k,i}}{{\tilde{P}}_k}---\left(19\right)$

Wherein with respectively desired bit rate and the mean bit rate for the user i in the k of community, and with respectively for the average power of the user i in the k of community with for the maximum downlink power limit of community k.Value can be interpreted as the speed of per unit of power, this value quantizes user's power efficiency.Thereby, value refer to reach the power that desired bit rate may need.Afterwards, by the needed power after gross power normalization, provided the relative value of intrasystem user's needed power contribution.

Notice that user may have a plurality of carryings.In this case, i is defined as to the index of the carrying in system more useful.In addition,, in practice, user or carrying may reach low-down bit rate, thereby cause the very high fluctuation of load.In order to overcome this problem, the alternate version of equation (19) is provided by following formula

${\mathrm{\&rho;}}_k=\underset{i}{\mathrm{\&Sigma;}}\min (C_i,\frac{{\tilde{R}}_{K,I}}{{\stackrel{\&OverBar;}{R}}_{k,i}})\frac{{\stackrel{\&OverBar;}{P}}_{K,I}}{{\tilde{P}}_k}---\left(20\right)$

Wherein, positive constant C _ibe used to bit rate ratio the upper limit is set, and reduce possible unsteadiness.

Thereby this paper describes a kind of method that allows the deployment Femto cell that power setting requires user and take into account.

Claims (47)

1. the method that control is given each delivering power of each subband in a plurality of subbands by the base station assigns of cellular communications networks, the method comprises:

Reception is from the information of at least one other base station of described network, and described information comprises that utilance function in the community about being served by described other base station is to the information to the susceptibility of the variation of the power of each subband by described base station assigns,

Identification improves the relatively useful subband of delivering power;

Whether the judgement factor relevant with the happiness of user in described community be over threshold value; And

Only, in the situation that the described happiness factor is less than described threshold value, just improve the delivering power in the subband identifying.

2. the method for claim 1, the average happiness of the user in the community of the wherein said happiness factor based on being served by described base station and being defined.

3. method as claimed in claim 2, the weighted average happiness of the user in the community of the wherein said happiness factor based on being served by described base station and being defined.

4. method as claimed in claim 2, the standard deviation of the happiness of the user in the community of the wherein said happiness factor based on being served by described base station and being defined.

5. method as claimed in claim 2, the percentile of the happiness of the user in the community of the wherein said happiness factor based on being served by described base station and being defined.

6. the method as described in claim 1-5, the ratio of the mean bit rate of the wherein said happiness factor based on each user and the needed bit rate of this user and being defined.

7. the method as described in claim 1-6, the wherein said happiness factor is based on introducing the scale factor of conservative degree and being defined.

8. method as claimed in claim 7, comprises that the current delivering power based in described community regulates described scale factor.

9. method as claimed in claim 7 or 8, comprises scale factor described in the measured interference adjustments based in described community.

10. a base station, is configured to carry out the method as described in the arbitrary claim in claim 1-9.

11. 1 kinds of methods that control is given each delivering power of each subband in a plurality of subbands by the base station assigns of cellular communications networks, the method comprises:

Information from the first base station is sent to at least one other base station of described network, described information comprises that utilance function in the community about being served by described the first base station is to the information to the susceptibility of the variation of the power of each subband by described other base station assigns

The step of wherein said transmission information comprises: by X2 interface, send information to described at least one other base station.

12. 1 kinds of methods that control is given each delivering power of each subband in a plurality of subbands by the base station assigns of cellular communications networks, the method comprises:

Information from the first base station is sent to at least one other base station of described network, described information comprises that utilance function in the community about being served by described the first base station is to the information to the susceptibility of the variation of the power of each subband by described other base station assigns

The step of wherein said transmission information comprises: transmit information relevant to the relative arrowband delivering power of described community in each subband in described subband.

13. methods as claimed in claim 12, wherein the described susceptibility of utilance function is determined based on described base station and the path gain between the corresponding transmitter at place, described other base station.

14. methods as described in claim 12 or 13, the power that wherein the described susceptibility of utilance function transmits in described subband based on described other base station and being determined.

15. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 11-14.

16. 1 kinds of methods of determining the disturbing effect in the communities served by the base station of cellular communications networks, the transmission of the base station of described interference at least one the contiguous community from described network causes, described method comprises:

From being connected to the mobile device of described base station, obtain measurement result; And

Utilize described measurement result obtain utilance function in the described community about being served by described base station to described other base station assigns measuring to the susceptibility of the variation of the power of each subband.

17. methods as claimed in claim 16, comprising:

In a plurality of subbands, obtain the measurement result from described mobile device.

18. methods as described in claim 16 or 17, comprising:

Repeat to obtain each measurement result in described measurement result, and for each the measurement result formation time mean value in described measurement result.

19. methods as described in claim 16-18, comprise for each subband in described subband and form average measurement result according to the measurement of being undertaken by a plurality of mobile devices.

20. methods as claimed in claim 19, comprise for each the average measurement result lower than threshold value, for the described susceptibility of the described utilance function of corresponding subband, are set to negative value.

21. methods as described in claim 16-20, wherein said measurement result is CQI measurement result.

22. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 16-21.

23. 1 kinds of methods of estimating the spectrum efficiency of the subband in the base station on cellular communications networks, described method comprises:

Power function by the CQI reported by the mobile device of measuring on described subband is similar to described spectrum efficiency;

By the linear function of the ratio of interference and noise being similar to described CQI by the measured signal of described mobile device, wherein said signal is measured by decibel to the ratio of interference and noise.

24. methods as claimed in claim 23, wherein said linear function comprises constant offset item.

25. methods as claimed in claim 24, wherein said spectrum efficiency ω is estimated as:

ω＝a(c′log ₁₀(γ)+d+Δd) ^b

Wherein, a, b, c ' and d are constants,

Δ d is constant offset item, and

γ is the ratio of signal to interference and noise.

26. methods as described in claim 24 or 25, also comprise based on measuring of channel quality regulated to described shift term.

27. methods as claimed in claim 26, the wherein said number that measuring of channel quality is comprised to HARQ request.

28. methods as claimed in claim 26, wherein saidly comprise bLock error rate to measuring of channel quality.

29. methods as described in claim 23-28, comprise and utilize estimated spectrum efficiency to obtain for the utilance function in the community of being served by described base station the value to the susceptibility of the variation of the power of each subband by another base station assigns.

30. methods as claimed in claim 29, comprise with respect to described signal the ratio of interference and noise are differentiated to estimated spectrum efficiency, to obtain the value of described susceptibility.

31. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 23-30.

32. 1 kinds of methods of controlling the base station in cellular communications networks, described method comprises:

For each user in a plurality of users, receive the value of the initial bit rate requirement that represents described user;

Determine and be required that the corresponding downlink power of distributing to described user is to reach corresponding bit rate requirement;

Determine that total downlink power requires the summation into desired described each downlink power; And

When total downlink power of described base station surpasses threshold value, the bit rate of at least one user in described user is required to be reduced to the value requiring lower than corresponding initial bit rate.

33. methods as claimed in claim 32, the step that wherein reduces described bit rate requirement comprises: threshold value is set, and described bit rate requires can not be lowered to below described threshold value.

34. methods as claimed in claim 33, wherein said threshold value is the predetermined score that described initial bit rate requires.

35. methods as claimed in claim 34, the type of service of wherein said predetermined score based on described user and being set up.

36. methods as described in the arbitrary claim in claim 32-35, wherein reduce the step that the bit rate at least one user in described user requires and comprise:

For a plurality of users, judge whether Available Bit Rate surpasses corresponding initial bit rate requirement;

Select at least one user, for this user, Available Bit Rate is lower than corresponding initial bit rate requirement; And

Reduction is for the bit rate requirement of selected at least one user in described user.

37. methods as claimed in claim 32, the step that wherein reduces described bit rate requirement comprises: reduce bit rate until average community utilance surpasses threshold value with respect to the size of the rate of change of average cell power.

38. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 32-37.

The method of the value of the load on 39. 1 kinds of base stations of calculating cellular communications networks, wherein said base station can utilize a plurality of subbands and can utilize frequency selectivity power to control, and described method comprises:

Average power based on for each user and mean bit rate calculate the value of described load.

40. methods as claimed in claim 39, the step of wherein calculating described load comprises: based on calculating the value of described load for the desired bit rate of each user and the ratio of bit rate that reaches and the average power of this user based on for each user and the ratio of maximum downlink power for community.

41. methods as claimed in claim 39, comprising:

The value of calculating described load is the summation with regard to all users, for the desired bit rate of each user and the ratio of the bit rate reaching and average power for this user with the product of the ratio of maximum downlink power for community.

42. methods as claimed in claim 41, also comprise:

When the ratio for the desired bit rate of described user and the bit rate that reaches is less than the corresponding predetermined value for one or more users, utilize corresponding predetermined value to replace calculating the value of described load for the desired bit rate of described user and the ratio of the bit rate reaching.

43. as the method as described in one of in claim 39-42, is included in and in access control and/or congestion control and/or load balance, uses calculated load value.

44. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 39-42.

45. 1 kinds of methods that control is given each delivering power of each subband in a plurality of subbands by the base station assigns of cellular communications networks, described method comprises:

In described base station, from being connected to the mobile device of described base station, obtain channel quality information;

For each subband, utilize the channel quality information from described mobile device to form average channel quality metric; And

According to described average channel quality metric, estimate about the utilance function in the community of being served by described base station the information to the susceptibility of the variation of the power of each subband by other base station assigns.

46. methods as claimed in claim 45, comprise each subband lower than threshold value for described average channel quality metric, and estimated susceptibility value is set to negative real number value.

47. 1 kinds of base stations, are configured to carry out the method as described in the arbitrary claim in claim 45-46.